Race Prediction Sheet
Predictions from your own results, and the honest caveats. The most useful column is the one where you write what actually happened.
BFCBrilliance · bfcbrilliance.com/tools/race-time-predictor
Your benchmark race
- Race
- Date
- Distance
- Time
- Conditions
- Felt like an honest effort?
Predictions — and what really happened
| Target race | Distance | Predicted | Target pace | Actual | Difference |
|---|---|---|---|---|---|
The formula
- Riegel (1977)
- T2 = T1 x (D2 / D1) ^ 1.06
- Units
- any — they cancel, as long as BOTH match
- Double the distance
- time x 2.085, not x 2
- Pace slows per doubling
- about 4.25%
- Valid range
- roughly 3.5 to 230 minutes
- Overall accuracy
- about 80% — 1 runner in 5 misses meaningfully
- ⚠ Assumes
- you are TRAINED for the target distance
Standard distances
- 5K
- 5 km / 3.107 miles
- 10K
- 10 km / 6.214 miles
- Half marathon
- 21.0975 km / 13.109 miles
- Marathon
- 42.195 km / 26.219 miles
Before believing a prediction
- Benchmark race was an honest, hard effort - not a training run
- Both distances entered in the SAME unit
- Distance gap is small - a 5K predicting a marathon is a guess
- You are actually TRAINED for the target distance
- Conditions on the day considered - heat, wind, hills, crowding
- Prediction sense-checked against your own history over two distances
- A slower fallback target set for a bad day
The formula does not know if you have trained
This is the assumption that catches everyone. Riegel describes how race performance SCALES with distance — it says nothing about whether you have done the long runs, practised eating on the move, or ever been on your feet for three hours. A 5K time predicts what a marathon-trained version of you could run. Whether that version exists yet is a separate question, and it is the one that decides your actual finish time.
Your own two results beat the exponent
1.06 is a population average. If you have raced two distances honestly, the exponent that fits YOUR two results describes you better than Riegel's does — and knowing whether you slow more or less than average is worth more than any single prediction.